Maxout_Network.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### **📈REPLICATING AND TESTING CUSTOM MAXOUT NETWORKS.**\n",
    "\n",
    "Hey there! This notebook is meant to compare the performance of 4 MLP models with different activation functions on the MNIST dataset.\n",
    "The first of these uses the typical ReLU activation, while the remaining three use various versions of **a custom maxout activation function** (replicating the original 2013 paper https://arxiv.org/abs/1302.4389) as defined in **Maxout.py**. \n",
    "\n",
    "The objective of this experiment is to see whether the theoretical benefits of Maxout work in real-world use cases. As outlined in the above paper, Maxout can (arbitrarily) approximate any continuous function as it acts much like a piecewise linear approximator - this is promising, as it means that both the ReLU and Leaky ReLU activation functions are both special cases of Maxout. Logically, this should mean that maxout can learn more complex functions and thus be able to better model an existing relationship.\n",
    "\n",
    "Practically, the function works by splitting the input vector into *i* groups of *k* neurons each. Then, it takes the **MAXIMUM** for each one of these groups - turning the group of *k* neurons into just one. So, if there are 10 neurons in total, and we split the layer into groups of 2 ($i=2$, $k=5$), then that means that activation function will map those 10 neurons **to just 2** - taking the **highest activations and capturing the relationships that appear to have the most relevance**.\n",
    "\n",
    "This essentially means that **maxout also has its own weights as it acts much like a linear transform* - this produces more trainable parameters but also offers the opportunity for better learning.\n",
    "\n",
    "**Mathematically:**\n",
    "$h_1(x)=max_{jΣ[1, k]}z_{ij}$, where $z_{ij}=x^TW_{...ij}+b_{ij}$\n",
    "\n",
    "**Visually:** (credit to @MlforNerds https://www.youtube.com/watch?v=DTVlyP-VihU&t=403s):\n",
    "\n",
    "<p align = \"center\"><img src=\"./images/MAXOUT_DIAGRAM_MLFORNERDS.jpg\" alt=\"Maxout Diagram\" /></p>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### **➗Why does Maxout actually work? (The Math)**\n",
    "\n",
    "As listed in the paper, there are ***two key principles*** that allow Maxout to (arbitrarily) approximate continuous functions. Consider the following neural network (from the original maxout paper):\n",
    "\n",
    "<p align = \"center\"><img src=\"./images/MAXOUT_NN_DIAGRAM.jpg\" alt=\"Maxout NN Example\" /></p>\n",
    "\n",
    "In this case, $v$ is simply the input vector of the previous layer. This is then split into two groups -> $z_1$ and $z_2$ (or $i_1$ and $i_2$ if using the notation from the above example). Each of these groups has a **single maxout activation**, specified by $h_1$ and $h_2$. The two outputs of these activations (and by extension the two groups) are fed into $g$, in essence converting $z_1$ and $z_2$ into a single output.\n",
    "\n",
    "Let's say, theoretically, that the weights of $g$ learned to **subtract the results from $h_2$ to $h_1$** - finding the difference between both functions. In theory, this last unit **should be able to approximate ANY continuous function**, as a result of:\n",
    "\n",
    "1. Two convex PWL (piecewise linear) functions **can be subtracted to generate ANY CONTINUOUS PWL FUNCTION**. Remember that Maxout basically approximates smooth functions in small \"pieces\". While this allows it to approximate convex functions, one Maxout unit alone cannot approximate any **non-convex PWL continuous function.** However, two of these convex PWL functions, when subtracted, can yield a continuous function -> meaning **that since $g(v) = h_1(v)-h_2(v)$ (since its a difference of two convex PWLs) it can approximate any continuous PWL function.**\n",
    "\n",
    "2. If we're trying to approximate a function, then *there exists a continuous PWL function that comes arbitrarily close* - in other words, if we choose an \"error\" $ε$ then there will be some PWL function that will yield less than $e$ when subtracted from the original function. More simply, $|f(v)-g(v)| < ε$ - meaning that **we can approximate any continuous function from any continuous PWL function with reasonable accuracy.**\n",
    "\n",
    "All in all, this means that the shown maxout network **can map any input/output relationship quite well, without problems!**"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### **📝Methodology + This Notebook.**\n",
    "Four models will be trained on Handwritten digit classification via MNIST (344k samples) - ReLU, Maxout with Bias, Maxout without Bias, and Maxout **with Dropout** (more on this in the next section).\n",
    "\n",
    "The ReLU model will be trained via FastAI, whereas the Maxout models will use FastAI *DataLoaders* and *Learners* but PyTorch for the training loop (as the activation function has been implemented separately). In the end, **all models are tested on the same 8 000 samples** and accuracy + confusion matrices are reported.\n",
    "\n",
    "Feel free to clone this repo and try it out for yourself! Reach out to aditya.dewan124@gmail.com, @adidewan124 on Twitter, or via LinkedIn if you have any questions or ideas."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 🎯**Results and Outcomes.**\n",
    "\n",
    "Interestingly, despite higher function-modelling capabilities, the Maxout network (both with and without biases and dropout) fell short by ~3% in terms of test accuracy when compared to the regular ReLU network.\n",
    "\n",
    "My hypothesis for this is that, despite having *higher potential to model functions, the excess parameters needed to be tuned by Maxout mean that they cause the model to be more overfitting-prone.* While the model achieves low training losses, the end accuracy on the test set **ends up being lower as excess parameters make it easier to adapt directly to the dataset rather than capture the underlying relationship behind it.** This explains why Maxout, as described by the paper, **is meant to be used in conjunction with Dropout** - more regularization is needed to unlock the function's full benefits.\n",
    "\n",
    "Surprisingly, however, the \"worst\" (by approximately 5 percentage points) performing model was the combination of Maxout and Dropout - in direct contradiction to the above hypothesis. I believe this is the case **as models that have dropout take significantly longer to learn the same relationships** - as parts of the network and activation functions are quite literally shut off, each part of the network becomes less prone to overfitting; this of course makes it more difficult for the network to learn relationships, but makes it more likely that the relationships learned are in better conjunction with reality.\n",
    "\n",
    "It appears that, despite being simpler, ReLU activations **allow the rest of the network to capture more meaningful relationships due to fewer parameters and increased simplicity!**\n",
    "\n",
    "**🔑 Key takeaway: Simplicity > Complexity for most cases** (Maxout with Dropout, while the theoretical best-solution, require more resources to achieve performance - but may be justified in cases where accuracy is paramount). \n",
    "\n",
    "*Special thanks to Ian Goodfellow, David Warde-Farley, Mehdi Mirza, Aaron Courville, and Yoshua Bengio for coming up with the original concept back in 2013! Check out their paper here https://arxiv.org/abs/1302.4389*"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2.6.3\n"
     ]
    }
   ],
   "source": [
    "#For datablocks and data processing + training\n",
    "import fastai\n",
    "print(fastai.__version__)\n",
    "from fastai.vision.all import *\n",
    "#For neural networks\n",
    "import torch.nn as nn\n",
    "#For activation functions and layers\n",
    "import torch.nn.functional as F\n",
    "#Import pandas and numpy\n",
    "import pandas as pd\n",
    "import numpy as np"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Import handwritten digit classification dataset via untar_data\n",
    "#Use a small subset of the MNISt dataset (344k samples)\n",
    "dataset = untar_data(URLs.MNIST)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Due to IPython and Windows limitation, python multiprocessing isn't available now.\n",
      "So `number_workers` is changed to 0 to avoid getting stuck\n",
      "torch.Size([64, 3, 28, 28])\n",
      "torch.Size([64])\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 648x648 with 9 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#Create a new datablock object (tells us preconditions for processing data)\n",
    "#We pass in the respective data to the block for processing\n",
    "datablock = DataBlock(\n",
    "            blocks = (ImageBlock, CategoryBlock),\n",
    "            get_items = get_image_files,\n",
    "            splitter = RandomSplitter(valid_pct = 0.2, seed = 32),\n",
    "            get_y = parent_label,\n",
    "            batch_tfms = aug_transforms(mult = 2.0, do_flip = False))\n",
    "\n",
    "#Create a data loader and pass in the training path -> will apply transforms and apply labels based on directories\n",
    "batch_size = 64\n",
    "training_dataloader = datablock.dataloaders(dataset/\"training\", batch_size = batch_size)\n",
    "training_dataloader.train.show_batch()\n",
    "sample_batch = training_dataloader.train.one_batch()\n",
    "\n",
    "print(sample_batch[0].shape)\n",
    "print(sample_batch[1].shape)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### DEFINING MODELS."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Create a ReLU and custom Maxout model and compare performance\n",
    "#We have three channels in our data - they are identical as the image is black and white\n",
    "n_channels = 3\n",
    "\n",
    "#ReLU Model (ReLU is a special case of Maxout with k = 2)\n",
    "ReLU_Model = nn.Sequential(\n",
    "    #Flatten inputs\n",
    "    nn.Flatten(),\n",
    "    #Basic Linear transform - shape of input images is 28*28 and has 3 channels; thus first input will have 28*28*3 samples\n",
    "    nn.Linear(28*28*n_channels, 500),\n",
    "    nn.ReLU(),\n",
    "    nn.Linear(500, 300),\n",
    "    nn.ReLU(),\n",
    "    nn.Linear(300, 200),\n",
    "    nn.ReLU(),\n",
    "    nn.Linear(200, 10),\n",
    "    #Use log softmax over softmax as gradient is easier to compute -> very close to the line x - max(x)\n",
    "    nn.LogSoftmax(dim=1)\n",
    ")\n",
    "\n",
    "#DEFINE CUSTOM MAXOUT MODELS\n",
    "from Maxout import CustomMaxout\n",
    "\n",
    "# WB - WITHOUT BIAS.\n",
    "# B - WITH BIAS.\n",
    "# D - WITH DROPOUT AND BIAS.\n",
    "\n",
    "#Keep everything the same but add maxout activations - set bias to FALSE\n",
    "Maxout_WB_Model = nn.Sequential(\n",
    "    #Flatten inputs\n",
    "    nn.Flatten(),\n",
    "    #Basic Linear transform - shape of input images is 28*28\n",
    "    nn.Linear(28*28*n_channels, 500),\n",
    "    #Custom Maxout - inputs, outputs, channels, bias\n",
    "    CustomMaxout(500, 400, n_channels, False),\n",
    "    nn.Linear(400, 300),\n",
    "    CustomMaxout(300, 200, n_channels, False),\n",
    "    nn.Linear(200, 100),\n",
    "    CustomMaxout(100, 50, n_channels, False),\n",
    "    nn.Linear(50, 10),\n",
    "    nn.LogSoftmax(dim=1)\n",
    ")\n",
    "\n",
    "#Declare another maxout model but with bias TRUE\n",
    "Maxout_B_Model = nn.Sequential(\n",
    "    #Flatten inputs\n",
    "    nn.Flatten(),\n",
    "    #Basic Linear transform - shape of input images is 28*28\n",
    "    nn.Linear(28*28*n_channels, 500),\n",
    "    #Same principal - but, bias is set to True (meaning that there are more parameters to tune)\n",
    "    CustomMaxout(500, 400, n_channels, True),\n",
    "    nn.Linear(400, 300),\n",
    "    CustomMaxout(300, 200, n_channels, True),\n",
    "    nn.Linear(200, 100),\n",
    "    CustomMaxout(100, 50, n_channels, True),\n",
    "    nn.Linear(50, 10),\n",
    "    nn.LogSoftmax(dim=1)\n",
    ")\n",
    "\n",
    "#Define dropout probability (0.2)\n",
    "drop_prob = 0.2\n",
    "Maxout_D_Model = nn.Sequential(\n",
    "    #Flatten inputs\n",
    "    nn.Flatten(),\n",
    "    nn.Linear(28*28*n_channels, 500),\n",
    "    CustomMaxout(500, 400, n_channels, True),\n",
    "    #Dropout Layer - as Maxout rapidly increases the parameters of a nn, increased regularization is needed to exceed baseline performance\n",
    "    #The Maxout paper itself recommends using Maxout in conjunction with Dropout for the best results; we'll be testing this\n",
    "    nn.Dropout(drop_prob),\n",
    "    nn.Linear(400, 300),\n",
    "    CustomMaxout(300, 200, n_channels, True),\n",
    "    #Include a dropout layer after every Maxout activation\n",
    "    nn.Dropout(drop_prob),\n",
    "    nn.Linear(200, 100),\n",
    "    CustomMaxout(100, 50, n_channels, True),\n",
    "    nn.Dropout(drop_prob),\n",
    "    nn.Linear(50, 10),\n",
    "    #Use log softmax over softmax as gradient is easier to compute -> very close to the line x - max(x)\n",
    "    nn.LogSoftmax(dim=1)\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "SuggestedLRs(valley=0.0006918309954926372)\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#Create learner for ReLU model\n",
    "ReLU_learner = Learner(training_dataloader, ReLU_Model, metrics = accuracy)\n",
    "#Graph learning rates - plot the minimum, steep, valley, and slide\n",
    "ReLU_optimalLR = ReLU_learner.lr_find()\n",
    "print(ReLU_optimalLR)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: left;\">\n",
       "      <th>epoch</th>\n",
       "      <th>train_loss</th>\n",
       "      <th>valid_loss</th>\n",
       "      <th>accuracy</th>\n",
       "      <th>time</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>0</td>\n",
       "      <td>0.415325</td>\n",
       "      <td>0.188314</td>\n",
       "      <td>0.941833</td>\n",
       "      <td>05:37</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: left;\">\n",
       "      <th>epoch</th>\n",
       "      <th>train_loss</th>\n",
       "      <th>valid_loss</th>\n",
       "      <th>accuracy</th>\n",
       "      <th>time</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>0</td>\n",
       "      <td>0.284557</td>\n",
       "      <td>0.117427</td>\n",
       "      <td>0.963083</td>\n",
       "      <td>05:37</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>1</td>\n",
       "      <td>0.281036</td>\n",
       "      <td>0.118861</td>\n",
       "      <td>0.962167</td>\n",
       "      <td>04:24</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2</td>\n",
       "      <td>0.249923</td>\n",
       "      <td>0.118378</td>\n",
       "      <td>0.964333</td>\n",
       "      <td>03:33</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>3</td>\n",
       "      <td>0.244258</td>\n",
       "      <td>0.099587</td>\n",
       "      <td>0.969917</td>\n",
       "      <td>03:32</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>4</td>\n",
       "      <td>0.228714</td>\n",
       "      <td>0.091881</td>\n",
       "      <td>0.973333</td>\n",
       "      <td>04:11</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>5</td>\n",
       "      <td>0.187598</td>\n",
       "      <td>0.080824</td>\n",
       "      <td>0.974667</td>\n",
       "      <td>04:16</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>6</td>\n",
       "      <td>0.162744</td>\n",
       "      <td>0.064349</td>\n",
       "      <td>0.981000</td>\n",
       "      <td>04:31</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>7</td>\n",
       "      <td>0.144693</td>\n",
       "      <td>0.059128</td>\n",
       "      <td>0.981917</td>\n",
       "      <td>03:51</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>8</td>\n",
       "      <td>0.138330</td>\n",
       "      <td>0.057244</td>\n",
       "      <td>0.982333</td>\n",
       "      <td>03:53</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>9</td>\n",
       "      <td>0.133055</td>\n",
       "      <td>0.055179</td>\n",
       "      <td>0.983333</td>\n",
       "      <td>04:34</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "#Extract optimal learning rate for relu model\n",
    "float_rlu_optimal_lr = ReLU_optimal_lr[0]\n",
    "#Fit and train model\n",
    "epochs = 10 \n",
    "#Callback at the end produces graph of loss\n",
    "ReLU_learner.fine_tune(epochs, float_rlu_optimal_lr, cbs = [ShowGraphCallback()])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Path('models/ReLU_Model.pth')"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Save model\n",
    "ReLU_learner.save(\"ReLU_Model\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### TRAINING MAXOUT MODELS - WITH BIAS, WITHOUT BIAS, AND DROPOUT (WITH BIAS)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "SuggestedLRs(valley=0.00019054606673307717)\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#Create learner object\n",
    "Maxout_WB_learner = Learner(training_dataloader, Maxout_WB_Model, metrics = \"accuracy\", cbs = [ShowGraphCallback()])\n",
    "#Find ideal learning rate\n",
    "Maxout_WB_optimalLR = Maxout_WB_learner.lr_find()\n",
    "print(Maxout_WB_optimalLR)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "SuggestedLRs(valley=0.00013182566908653826)\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#Create learner object\n",
    "Maxout_B_learner = Learner(training_dataloader, Maxout_B_Model, metrics = \"accuracy\", cbs = [ShowGraphCallback()])\n",
    "#Find ideal learning rate\n",
    "Maxout_B_optimalLR = Maxout_B_learner.lr_find()\n",
    "print(Maxout_B_optimalLR)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "SuggestedLRs(valley=0.0002290867705596611)\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#Find learning rate for dropout model\n",
    "Maxout_D_learner = Learner(training_dataloader, Maxout_D_Model, metrics = \"accuracy\", cbs = [ShowGraphCallback()])\n",
    "#Find ideal learning rate\n",
    "Maxout_D_optimalLR = Maxout_D_learner.lr_find()\n",
    "print(Maxout_D_optimalLR)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Begin training fastai learner\n",
    "#One of the key benefits of fastai is customization. As we have a custom activation function, some Recorder functionality in the fastai library does not work as intended.\n",
    "#To solve this, we can write our own training loop in native PyTorch.\n",
    "\n",
    "#Import matplotlib to display training graph and update as epochs progress\n",
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline\n",
    "\n",
    "#List of all models - all other variable names are derived from these models (i.e. (model1)_learner, (model1)_y_true)\n",
    "model_list = [\"ReLU\", \"Maxout_WB\", \"Maxout_B\", \"Maxout_D\"]\n",
    "\n",
    "#Define a) learning rate as a float (FastAI returns it as a 1 dim tensor) and b) declare optimizer\n",
    "def defineParams(model):\n",
    "    globals()[model + \"optimalLR_float\"] = globals()[model + \"_optimalLR\"][0]\n",
    "    #Use ADAM optimizer for consistency with ReLU model\n",
    "    globals()[model + \"_learner\"].opt = Adam(globals()[model + \"_learner\"].parameters(), globals()[model + \"optimalLR_float\"])\n",
    "#Define parameters\n",
    "[defineParams(m) for m in model_list[1:]]\n",
    "\n",
    "#Train model function\n",
    "def train_model(learner, n_epochs, train_iters, valid_iters):\n",
    "    #Train losses\n",
    "    train_losses = []\n",
    "    #Validation losses\n",
    "    valid_losses = []\n",
    "    #Array of epochs (for plotting the graph at the end of training)\n",
    "    epochs = []\n",
    "\n",
    "    #Iterate over epochs\n",
    "    for epoch in range(n_epochs):\n",
    "        #Iterate over each \"minibatch\" in batch (each sample in batch, batch size declared while we were using the dataloader)\n",
    "        #This is for TRAINING data\n",
    "        train_loop(learner, train_iters, train_losses, epoch, epochs)\n",
    "        #Same process but with validation data and WITHOUT computing gradients\n",
    "        valid_loop(learner, valid_iters, valid_losses)\n",
    "    \n",
    "    #If we are on the last epoch, return the average loss for the entire training session and create a graph of training and validation losses\n",
    "    if epoch == n_epochs - 1:\n",
    "        #Change figure size\n",
    "        plt.figure(figsize = (8, 6))\n",
    "        #Plot train and validation losses\n",
    "        plt.plot(epochs, train_losses, label = \"TRAIN\")\n",
    "        plt.plot(epochs, valid_losses, label = \"VALIDATION\")\n",
    "        #Set titles\n",
    "        plt.xlabel(\"EPOCH\")\n",
    "        plt.ylabel('LOSS')\n",
    "        plt.title(\"MAXOUT MODEL - TRAINING AND VALIDATION LOSSES.\")\n",
    "        #Create a legend and show graph\n",
    "        plt.legend()\n",
    "        plt.show()\n",
    "\n",
    "#Train loop for a single epoch\n",
    "def train_loop(learner, train_iters, train_losses, epoch, epochs):\n",
    "    #Set training epoch loss to zero\n",
    "    train_epoch_loss = 0.0\n",
    "    #Iterate over training data\n",
    "    for batch_index, batch_data in enumerate(learner.dls.train, 0):\n",
    "        #Get data inputs and labels\n",
    "        data_inputs, data_labels = batch_data\n",
    "        #Set gradient to zero\n",
    "        learner.zero_grad()\n",
    "        #Make predictions (learner.model)\n",
    "        preds = learner.model(data_inputs)\n",
    "        #Calculate loss\n",
    "        calc_loss = learner.loss_func(preds, data_labels)\n",
    "        #Calculate gradients\n",
    "        calc_loss.backward()\n",
    "        #Use gradients to take a \"step\" with the optimizer - optimizer is, by default, set to Adam\n",
    "        learner.opt.step()\n",
    "        #Append loss - .item() converts scalar to regular float\n",
    "        train_epoch_loss += calc_loss.item()\n",
    "        #Check if we are on the last iteration of the minibatch\n",
    "        if batch_index == train_iters - 1:\n",
    "            avg_epoch_loss = train_epoch_loss/train_iters\n",
    "            train_losses.append(avg_epoch_loss)\n",
    "            epochs.append(epoch + 1)\n",
    "            print(\"EPOCH [{}] TRAINING LOSS: {}\".format(epoch + 1, avg_epoch_loss))\n",
    "            #Stop the loop \n",
    "            break\n",
    "\n",
    "#Validation loop for a single epoch\n",
    "def valid_loop(learner, valid_iters, valid_losses):\n",
    "    #Set validation loss to zero\n",
    "    valid_epoch_loss = 0.0\n",
    "    #Iterate over validation samples\n",
    "    for valid_batch_index, valid_batch_data in enumerate(learner.dls.valid, 0):\n",
    "        #Get x inputs and y labels\n",
    "        valid_inputs, valid_labels = valid_batch_data\n",
    "        #Get predictions\n",
    "        val_preds = learner.model(valid_inputs)\n",
    "        #Calculate validation loss\n",
    "        val_loss = learner.loss_func(val_preds, valid_labels)\n",
    "        #Append loss\n",
    "        valid_epoch_loss += val_loss.item()\n",
    "        #If on the last iteration of minibatch, print loss and save best model\n",
    "        if valid_batch_index == valid_iters - 1:\n",
    "            avg_val_loss = valid_epoch_loss/valid_iters\n",
    "            valid_losses.append(avg_val_loss)\n",
    "            print(\"VALIDATION LOSS: {}\".format(avg_val_loss))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "EPOCH [1] TRAINING LOSS: 0.7292311487992604\n",
      "VALIDATION LOSS: 0.28682384729385374\n",
      "EPOCH [2] TRAINING LOSS: 0.4730431707302729\n",
      "VALIDATION LOSS: 0.19886194705963134\n",
      "EPOCH [3] TRAINING LOSS: 0.39393509193261467\n",
      "VALIDATION LOSS: 0.1840494902431965\n",
      "EPOCH [4] TRAINING LOSS: 0.35080582402149835\n",
      "VALIDATION LOSS: 0.14306614715605975\n",
      "EPOCH [5] TRAINING LOSS: 0.3145756204823653\n",
      "VALIDATION LOSS: 0.1571316672116518\n",
      "EPOCH [6] TRAINING LOSS: 0.2932680302460988\n",
      "VALIDATION LOSS: 0.13387144025415182\n",
      "EPOCH [7] TRAINING LOSS: 0.2723684365351995\n",
      "VALIDATION LOSS: 0.12167629271745682\n",
      "EPOCH [8] TRAINING LOSS: 0.2621143926580747\n",
      "VALIDATION LOSS: 0.13855159293860198\n",
      "EPOCH [9] TRAINING LOSS: 0.24829426196217536\n",
      "VALIDATION LOSS: 0.10243323680013418\n",
      "EPOCH [10] TRAINING LOSS: 0.24022699278096357\n",
      "VALIDATION LOSS: 0.09487039815634489\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 576x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#Train the Maxout models (without bias)\n",
    "#Total epochs will be 10, and training iterations will be 750 and 50 respectively (using the same # of iterations as the ReLU model)\n",
    "train_model(learner = Maxout_WB_learner, n_epochs = 10, train_iters = 750, valid_iters = 50)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Path('models/Maxout_WB_Model.pth')"
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Save Maxout model (without bias)\n",
    "Maxout_WB_learner.save(\"Maxout_WB_Model\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "EPOCH [1] TRAINING LOSS: 1.2623714292844137\n",
      "VALIDATION LOSS: 0.44069181084632875\n",
      "EPOCH [2] TRAINING LOSS: 0.6384548414150873\n",
      "VALIDATION LOSS: 0.2779819968342781\n",
      "EPOCH [3] TRAINING LOSS: 0.4851772368748983\n",
      "VALIDATION LOSS: 0.20500323802232742\n",
      "EPOCH [4] TRAINING LOSS: 0.4092070794502894\n",
      "VALIDATION LOSS: 0.17630425751209258\n",
      "EPOCH [5] TRAINING LOSS: 0.36215244643886885\n",
      "VALIDATION LOSS: 0.18197616711258888\n",
      "EPOCH [6] TRAINING LOSS: 0.3282258113423983\n",
      "VALIDATION LOSS: 0.16159541741013528\n",
      "EPOCH [7] TRAINING LOSS: 0.3002926664153735\n",
      "VALIDATION LOSS: 0.13213765066117047\n",
      "EPOCH [8] TRAINING LOSS: 0.28222837359209857\n",
      "VALIDATION LOSS: 0.12845431409776212\n",
      "EPOCH [9] TRAINING LOSS: 0.2638839848836263\n",
      "VALIDATION LOSS: 0.11626705270260572\n",
      "EPOCH [10] TRAINING LOSS: 0.25153289519747096\n",
      "VALIDATION LOSS: 0.10882409835234284\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 576x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#Train the Maxout models (with bias)\n",
    "#Total epochs will be 10, and training iterations will be 750 and 50 respectively (using the same # of iterations as the ReLU model)\n",
    "train_model(learner = Maxout_B_learner, n_epochs = 10, train_iters = 750, valid_iters = 50)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Path('models/Maxout_B_Model.pth')"
      ]
     },
     "execution_count": 29,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Save Maxout model (with bias)\n",
    "Maxout_B_learner.save(\"Maxout_B_Model\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "EPOCH [1] TRAINING LOSS: 1.16579405216376\n",
      "VALIDATION LOSS: 0.4077025577425957\n",
      "EPOCH [2] TRAINING LOSS: 0.6055373674035073\n",
      "VALIDATION LOSS: 0.2531306692957878\n",
      "EPOCH [3] TRAINING LOSS: 0.484522975564003\n",
      "VALIDATION LOSS: 0.2269952890276909\n",
      "EPOCH [4] TRAINING LOSS: 0.42294044329722724\n",
      "VALIDATION LOSS: 0.19595682010054588\n",
      "EPOCH [5] TRAINING LOSS: 0.38213305350144705\n",
      "VALIDATION LOSS: 0.17838216871023177\n",
      "EPOCH [6] TRAINING LOSS: 0.3574310518105825\n",
      "VALIDATION LOSS: 0.18633492290973663\n",
      "EPOCH [7] TRAINING LOSS: 0.332657618890206\n",
      "VALIDATION LOSS: 0.16271564692258836\n",
      "EPOCH [8] TRAINING LOSS: 0.3101572586596012\n",
      "VALIDATION LOSS: 0.13308388751000166\n",
      "EPOCH [9] TRAINING LOSS: 0.30444542261461416\n",
      "VALIDATION LOSS: 0.13548599317669868\n",
      "EPOCH [10] TRAINING LOSS: 0.2924726982712746\n",
      "VALIDATION LOSS: 0.13778638917952776\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAfEAAAGDCAYAAAA72Cm3AAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjUuMSwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/YYfK9AAAACXBIWXMAAAsTAAALEwEAmpwYAABEkklEQVR4nO3deXxcdb3/8dcn+9qkTaZb0r1NKZQuUMreVBAEBNmhVVkULj+vIvfq/enFBS/ei/eqV70K4kV+oEWWlh3LJioqBQrSlrZAW7pv6Zqke9I02/f3xzlJJtPJ1mZyZpL38/GYR2bOOXPOZyaZvOf7Ped8jznnEBERkcSTFHQBIiIicmwU4iIiIglKIS4iIpKgFOIiIiIJSiEuIiKSoBTiIiIiCUohLiIikqAU4gnAzDaZWa2ZFUZMX2ZmzsxGRky/258+PWL6V83sIzNLC5v2z2a21MxS/Mc3m9mHZlZtZjvN7H/NLD9s+Tlmdk/Eekf620sxs0Nht0YzOxz2+HNRXtsc/7mfiZj+c3/6zWHTis3scTOrNLMqM3vPzC6NeJ7z5x3yl3vdzK6PWOZvZlYTUeuL/ryZZlYW9RfRBWb2QNi6a82sLuzxq2HvWdO0TWZ2Z5T1zPSX+2bE9Ob3POJ9nB62zFgzc2GP/2Zmt0as9/6I9b4V8Z4PMbP/Z2bb/To3+Ns6oYPXP8r//f8qyjzn/40lhU27x8zmRLy2pvdml5m9ZGYXtLO9j83si1Gm/5OZLQ57fLO/7usilmvz9x7+N9+V2sws21/mlbBpK8Ke2xDxd/htv763ItbT0Wey6fN+bdi0FIvyvyFsfvPfQpR5J5rZfDPbb2YHzeyvZnZWxDK3+O/5Qf89eNnMcv15xWb2rJlV+Ov4sOlvKsr713S7vqPnSnQK8cSxEZjd9MDMTgYyIxcyMwNuAPYAN0XMvh/YB3zHX3Y08H3gFudcvZn9C/Aj4BtAHnAGMAL4k4UFf3ucczlNN2ALcFnYtMfbeNqa8Fr9YLoWWB82bQDwFlALnAQUAv8DPGFm10Ssb7K//fHAHOCXZvZvEcvcHl6rc+6yzry+znLOfSnsffhP4MmwbV0ctmi+v8w1wF1RwuAmov8uo9kD3NPhUi2qgBvb+UdfACwEsoBzgVzgFOANoM1A9d0I7AVmmVl6lPlDgVkdrKPpvZkM/Al4vp1/6I/424x0gz+vSVfez+Ot7RrgCHChmQ0BcM6dFPZ38Sat/w7/M3IjXfhM7gH+3cySj+dFmdkY4G3gQ2AU3u/peeCPZnamv0wp3t/0bOdcLjABeCpsNY8CW/06C/B+L7siNpUf8fl7sgvPlTAK8cTxKK3/Sd0E/C7KcufiffD+Ce8faPMH3TnXCNwCfM3MJgH/D/iVc+59M+uHF+hfdc79wTlX55zbBFyH94H6fAxeU5MXgbPNrL//+CLgA2Bn2DJfAw7hfeHY6Zw77JybC/wA+Kn/5aUV51yFc+5R4B+Bb/mhFJecc4uBFcCUpmlmloUXBF8BxpnZtA5W8wgwyf8n2xn78L7kRH7BafI14ABwg3NuvfPsc8791jl3XwfrvhH4LlAHRPuC9GPg+/4Xtnb5v+9fAHcDPwpvwYd5FDjHzEY0TTCzCcAkYK7/eARQCtwGfMrMBnW07eOs7SbgAby/5aN6oTrSxc/kH/C+4B7v5/Ru4B3n3Hecc3uccwedc/fivb8/8pc5zV9mKYC/3CPOuYNh8+c456qcc/XOuaXOuVc7uf3jeW6fpBBPHO8C/cxsgv9t+3rgsSjL3YQXik3fbFt1NzvnVgP/BfwFKMb7JwFwFpABPBex/CHgVTpueR2PGmA+LS2zGzn6C8oFwLP+F5FwTwHDgZJ21v97IAWY3s4ygTKzM4CJwLqwyVfjfXF5GniN6C3NcNV4LaQfdGHTPwCuNrPxUeZ9Eng+ynveLjM7F+9vax7e7yda3c/hfUG4uQurfg4YiNfD0opzrgz4K17Lu8mNwCvOuYqwx4udc88CqziGYO1sbWY2HJgJPO7fOvrdRdOVz6QD7gL+zcxSj2FbTS7A+3uL9BTeF+0s4O94X4K+b2ZnR+lpeRe438xm+e9DVxzPc/skhXhiaWqNXwB8DGwLn+l/wK4FnnDO1QHPEL3b8E28rqpnnHM1/rRCoMI5Vx9l+R3+/Fj6HV7Xbh5ea+mFiPmFfh2RdoTNj8p/LyqAAWGT7zWzfWG3/zjmyo9PhZkdBt4BfkXr130TXjd8A/AEMLsT/6B/DQw3s4s7WA7wWpJ4rcV/jzK7kLDeEDP7jP9eHTSzP7az2puAV51ze/26LzazgZGbxgud77XR3R7Ndv/ngDbmP4If4n6L+HO07kq/0a8H/+fxdqm3V9uNwAfOuZV4PQEnmdnULq6zS59J59x8oByIuq+7C9ts63OWBPR3zr0JXIW3a+VloNLMfhbWlX8t3v+Yu4CN5h27c1rE+ioiPn8TuvBcCaMQTyyPAp/Fa71E60q/EqgHmg6keRzvH2ioaQG/e/3XwH3A7f5+cfBCrrCN7s0h/nz89UcGSSrQ6N+OiXPuLSCE1wX7knPucMQiFX4d0Wprmh+VH3whvP2GTe5wzuWH3e7qSr1mdm7YQTkruvLcCIVADvB/8Vpuqf76hwGfwPsdgtebkAF8ur2VOeeOAP/h347axdCGH+G1rCZHTK8k7D13zs13zuXjdbNHPUbCzDLx/hE/7j/nHbxjIz4bpdZX/Hm3dbLOIv/nnjbmPwcM8Xs1ZuLty3/Zr+tsvH288/xlnwBONrMpndx2V2u7kZb3YDvecQRd/dLQ2c9kuO/iHfOS0cVthW+zrc9ZI95xDjjnXvWPIxkAXI73P+lWf95e59ydzrmTgEHAMuCFiF1ehRGfv1VdeK6EUYgnEOfcZrwD3C4hoovNdxNeIGwxs5143WKphB0Qh/cNdzfePvMH8AIdvJbgEbxv2M3MLBu4GHjdn7QFGBmx3VHA1q52u0bxGPAvRP+C8me8bt/Iv9nr8A6EWdPOei/H+/Lx3nHW18w592bYQTknHee6GpxzP8XbrfBlf/INeJ/PF/3f5Qa8f8yd6Zb9Ld5BUFd2cvuVwM/xgj/c68AVbeyDbsuVQD/gV+YdSb0TL+DaqrspdLI6ue7dwOpoM51z1Xi9TzfivX/znHO1/uyb8L7ULPNr+rs//Vi6udutzbwjucfhHYfR9B6cjteT0uExAGE6+5ls5pz7E94umS9HzuukP+N9CYt0Hd5+8OqI7TU6517H2z03MUo9FcBP8I7TaasHJarjeW5fohBPPLcA5znnqsInmlkRcD7ePvAp/m0yXivrJn+ZycAdwD845xzeQSwjzewLzrn9ePvH7zOzi8ws1T9q+WmgDK8XAOBZ4NNmdqGZJZvZULx/xE0tnONxL96uggVR5v0PXjg8bGaDzSzDzGbjBcA3/NfTipkNMO+0tvuBH/lh1Sn++sNvPdES+CHwTTNrCuvv0/K7nIK3j/zTHR2g53e/3g38axe2/TO8fbATIqb1Bx41szHmySXs4LsobgJ+A5wcVvfZwBTzzqiIrPVveEdCt9lKNbNBZnY73gF43+rgy+IjeMeLXO3fx38/r8Nr8U8Ju30V+Fx4sHb1995GbTfhHbF+Yti2JuJ9UenUbg6ALnwmI30H+GYb88KlRLzWVH97Z5nZD/zPT66ZfRXv7/Ff/dd8ub/Pur//NzEdbxfYu/78H5nZRPNOc8vFO7B0XWc+f8fz3D7LOadbnN+ATcAno0xPwdu3OBK4E1gSZZmheEcITwYWA9+MmD8TrwttkP/4FuAj4DDeqR2/xtsPFv6cy4AlwH5gM/DfQGZn645YZg5wTxvz3gJuDns8HG//4h6806MWAZdHPMf58w75y/0V+GzEMn/Da/UeCrstCXs/XJTb2OP4/d0NPBYxbaS/3pSwaYZ3hPpdfn2hKOtaAdwe+fzI9xHvC/pH3ke81eu+Nex1lkWs+5v+OsPf86HAw3j7RA/hnfb3CDAhSm1FeD0eJ0eZ9wrwk7Df0diweaf70+ZEvDeH/N/lbv/5F3XivTa8XotVYdNm+fWnRiybgfe3f2l7v/fw97aj2vx17sU7tTKytl/hHYdy1O8jbNrNwFsR09r9TLbx9/WKX+fINt6nv0V5rY/58yYCL+EdeHjIX/acsOfOwOsFqAAO4vWCfTNs/n3AWv+55f66JkR5/8JvX+/ouf78Q8C5x/pZ7I03898YERERSTDqThcREUlQCnEREZEEpRAXERFJUApxERGRBKUQFxERSVBdGXggLhQWFrqRI0cGXYaIiEiPWbJkSYVzLhQ5PWYhbma/wTv/crdz7qiRfPxBOJoGozgE/KNzbnlH6x05ciSLFy/uaDEREZFew8w2R5sey+70OXiXlGzLRqDUOTcJb7jHB2NYi4iISK8Ts5a4c26BP0RgW/MXhj18F+/ShSIiItJJ8XJg2y1418cVERGRTgr8wDYz+wReiJ/TzjK34V+ucPhwXSdeRCSW6urqKCsro6amJuhS+pyMjAyKi4tJTY284nN0gYa4mU0CHgIudu1cpcY59yD+PvNp06ZpsHcRkRgqKysjNzeXkSNHokt59xznHJWVlZSVlTFq1KhOPSew7nQzG453TewbnHPtXQtaRER6UE1NDQUFBQrwHmZmFBQUdKkHJJanmM3Fu7xfoZmV4V1vNxXAOfcA8D2gAPiV/4dS75ybFqt6RESk8xTgwejq+x6zlrhzbrZzbohzLtU5V+yce9g594Af4DjnbnXO9XfOTfFvCnAREaGyspIpU6YwZcoUBg8eTFFRUfNjM2PKlClMnDiRyy67jH379rV67uTJk5k9e3araTfffDPPPPMMADNnzmTatJa4Wbx4MTNnzoz1S4qZeDk6XUREBICCggKWLVvGsmXL+NKXvsTXvva15sfZ2dksW7aMjz76iAEDBnD//fc3P2/VqlU0NjayYMECqqqq2lz/7t27efXV3nFClEJcREQS0plnnsm2bduaHz/xxBPccMMNXHjhhcyfP7/N533jG9/gnnvu6YkSYy7wU8xERCR+ff/FFazcfqBb13ni0H7822UnHdc6GhoaeP3117nllluapz355JP86U9/YvXq1fzyl788qlu9yZlnnsnzzz/PX//6V3Jzc4+rjqD16Zb4oSP1/OGjHdTUNQRdioiIdMLhw4eZMmUKBQUF7NmzhwsuuACARYsWEQqFGDFiBOeffz7vv/8+e/fubXM93/3ud3tFa7xPt8QXb9rDlx57nzlfOI2Z4wcGXY6ISNw53hZzd8vMzGTZsmXs37+fSy+9lPvvv5877riDuXPn8vHHH9N0lcsDBw7w7LPPcuutt0Zdz3nnncddd93Fu+++24PVd78+3RI/Y3QB6SlJLFhTEXQpIiLSBXl5edx777385Cc/4ciRIzz99NN88MEHbNq0iU2bNvH73/+euXPntruO73znO/z4xz/uoYpjo0+HeEZqMqePLuCNNbuDLkVERLpo6tSpTJ48maeeeoqioiKKioqa582YMYOVK1eyY8eONp9/ySWXEAoddYnuhGLOJdYoptOmTXPdeT3xh97cwD0vr+LNb36CYQOyum29IiKJatWqVUyYMCHoMvqsaO+/mS2JNp5Kn26JA8wc730LW7C2POBKREREuqbPh/iYUA5F+ZksWKMQFxGRxNLnQ9zMmFES4u11ldQ1NAZdjoiISKf1+RAHKC0p5NCRet7f3PY5hSIiIvFGIQ6cNbaQ5CTjDXWpi4hIAlGIA/0yUjl1eH8d3CYiIglFIe4rHR/io20HKD94JOhSRET6tJkzZ/Laa6+1mvbzn/+cL3/5y5SXl5Oamsqvf/3rVvNHjhxJRUXrgbvmzJnD7bffDsDdd9/dfEnTcePGcdVVV7Fy5cpWyy9duhQza972lVdeyZQpUxg7dix5eXnNl0NduHAhM2fOpOl05/3793PjjTcyZswYxowZw4033sj+/fsB2LRpE2bGfffd17yd22+/nTlz5hz/G4VCvFlpiXeq2ZtqjYuIBGr27NnMmzev1bR58+Yxe/Zsnn76ac4444wOR2OLpumSpmvXruX666/nvPPOo7y85X/+3LlzOeecc5rX/fzzz7Ns2TIeeughzj333ObLoZ511lmt1nvLLbcwevRo1q9fz/r16xk1alSr4V4HDhzIL37xC2pra7tcc0cU4r4Th/SjIDtN+8VFRAJ2zTXX8NJLL3HkiNczumnTJrZv394csD/96U8pKytrdRnSrrr++uu58MILeeKJJwBwzvHMM88wZ84c/vjHP1JTU9Op9axbt44lS5Zw1113NU/73ve+x+LFi1m/fj0AoVCI888/n0ceeeSY621Ln74ASrikJO9UszfWlNPY6EhKsqBLEhEJ3qt3ws4Pu3edg0+Gi3/Y5uyCggKmT5/OH/7wBy6//HLmzZvH9ddfT1lZGTt37mT69Olcd911PPnkk3z9618/5jJOOeUUPv74YwDefvttRo0axZgxY5g5cyavvPIKV111VYfrWLlyJVOmTCE5Obl5WnJyMlOmTGHFihVMmjQJgDvvvJOLL76YL37xi8dcbzRqiYcpLQmxp6qWj7bvD7oUEZE+LbxLvakrfd68eVx33XUAzJo165i61MOFDzs+d+5cZs2a1eV1O+cwO7rRFzl91KhRTJ8+vbnl313UEg9z7rhCzOCN1eVMKs4PuhwRkeC102KOpSuuuIKvf/3rvP/++xw+fJhTTjmFW2+9lV27dvH4448DsH37dtauXcu4ceOOaRtLly5l2rRpNDQ08OyzzzJ//nx+8IMf4JyjsrKSgwcPkpub2+46TjrpJJYuXUpjYyNJSV67uLGxkeXLlx81/vm3v/1trrnmGmbMmHFM9UajlniYgpx0Jg7N035xEZGA5eTkMHPmTL74xS8ye/ZsVq9eTVVVFdu2bWu+3Oi3vvWtow6A66xnn32WP/7xj8yePZs///nPTJ48ma1bt7Jp0yY2b97M1VdfzQsvvNDhesaOHcvUqVO55557mqfdc889nHLKKYwdO7bVsieccAInnngiL7300jHVHI1CPEJpSYilW/ex/3Bd0KWIiPRps2fPZvny5c3d21deeWWr+VdffXWrbu9JkyZRXFxMcXFx1H3l//M//9N8itljjz3GX/7yF0KhUJvr7mzX98MPP8yaNWsYO3YsY8aMYc2aNTz88MNRl/3Od75DWVlZp9bbGX3+UqSRFm3aw7UPvMP/fu4ULj55SMy2IyISr3Qp0mDpUqTHYeqwfHIzUtSlLiIicU8hHiElOYmzxxTyxppyEq2XQkRE+haFeBSl40Ps2F/D2t2Hgi5FRESkTQrxKGb4Q7AuUJe6iPRR6okMRlffd4V4FEX5mYwbmKP94iLSJ2VkZFBZWakg72FN56dnZGR0+jka7KUNM0pCPPrOZqpr68lK09skIn1HcXExZWVlrS4OIj0jIyOD4uLiTi+vdGpDaUmIh9/ayN837OETJwwMuhwRkR6TmprKqFGjgi5DOkHd6W2YPmoAGalJ6lIXEZG4pRBvQ0ZqMmeMLtDBbSIiErcU4u2YMS7EhooqtlRWB12KiIjIURTi7Sgd751q9sZatcZFRCT+KMTbMbowm+L+mepSFxGRuKQQb4eZUVoSYuG6CmrrG4MuR0REpBWFeAdmlISoqm1gyea9QZciIiLSikK8A2eNKSAlyXSqmYiIxB2FeAdyM1I5dUR/7RcXEZG4oxDvhNLxIVbuOMDuAzVBlyIiItJMId4JpU1XNVtbEXAlIiIiLRTinTBhcD8Kc9K1X1xEROKKQrwTkpKMGSWFvLW2nIZGXZpPRETig0K8k0pLQuytruPDbfuDLkVERARQiHfaueNCmMEbq9WlLiIi8UEh3kkDstOYVJTHG2t2B12KiIgIoBDvktKSEMu27mN/dV3QpYiIiCjEu6J0fIhGB2+t06lmIiISPIV4F0wuzqdfRoq61EVEJC7ELMTN7DdmttvMPmpjvpnZvWa2zsw+MLNTYlVLd0lJTuKccYW8saYc53SqmYiIBCuWLfE5wEXtzL8YGOffbgP+N4a1dJvSkhC7Dhxh9a6DQZciIiJ9XMxC3Dm3ANjTziKXA79znneBfDMbEqt6usuMpiFYNXqbiIgELMh94kXA1rDHZf60o5jZbWa22MwWl5cHG55D8jIZPyhXQ7CKiEjgggxxizIt6o5m59yDzrlpzrlpoVAoxmV1bEZJIYs27qXqSH3QpYiISB8WZIiXAcPCHhcD2wOqpUtKSwZS29DIuxsqgy5FRET6sCBDfD5wo3+U+hnAfufcjgDr6bRpI/uTmZqs/eIiIhKolFit2MzmAjOBQjMrA/4NSAVwzj0AvAJcAqwDqoEvxKqW7paRmsyZYwq0X1xERAIVsxB3zs3uYL4DvhKr7cfajHGF/OXj3WyqqGJkYXbQ5YiISB+kEduOUen4gQAsWKvWuIiIBEMhfoxGFmQxfECW9ouLiEhgFOLHyMwoLQmxcH0lR+obgi5HRET6IIX4cZhREqK6toElm/YGXYqIiPRBCvHjcOaYAlKTTUepi4hIIBTixyEnPYVpIwYoxEVEJBAK8eNUOj7ExzsPsutATdCliIhIH6MQP04zxnljuas1LiIiPU0hfpwmDMkllJuuEBcRkR6nED9OTaeavbW2gobGqBdhExERiQmFeDcoLQmx/3Ady8v2BV2KiIj0IQrxbnDO2EKSDN5YrS51ERHpOQrxbtA/O41JxfnaLy4iIj1KId5NSktCfFC2j71VtUGXIiIifYRCvJuUjg/R6OCtdRVBlyIiIn2EQrybTC7OJy8zVV3qIiLSYxTi3SQ5yThnXCFvrCnHOZ1qJiIisacQ70alJSHKDx5h1Y6DQZciIiJ9gEK8G5WWeEOwLlirLnUREYk9hXg3GtQvgxMG5+p8cRER6REK8W5WWhJi8eY9HDpSH3QpIiLSyynEu1lpSYi6Bsc76yuDLkVERHo5hXg3O3Vkf7LSklmgU81ERCTGFOLdLD0lmbPGFPC3Nbt1qpmIiMSUQjwGZpSE2LrnMJsqq4MuRUREejGFeAw0nWr2xurdAVciIiK9mUI8BkYUZDOyIIsFazWOuoiIxI5CPEZKS0K8s76SmrqGoEsREZFeSiEeIzNKQhyua2Dxpr1BlyIiIr2UQjxGzhhdQFpyEm+s0X5xERGJDYV4jGSnp3DaqP4sWKP94iIiEhsK8RgqLQmxetdBduw/HHQpIiLSCynEY2hG01XNNHqbiIjEgEI8hsYPymVQv3TeUIiLiEgMKMRjyMwoLQnx1toK6hsagy5HRER6GYV4jJWWDORATT3Ly/YFXYqIiPQyCvEYO2dsIUkGb6xWl7qIiHQvhXiM5WWlMmVYvvaLi4hIt1OI94DSkoF8sG0/e6pqgy5FRER6EYV4DygdH8I5eHOtWuMiItJ9FOI94OSiPPpnpapLXUREupVCvAckJxnnjAuxYE0FjY0u6HJERKSXUIj3kNKSEBWHjrByx4GgSxERkV5CId5DZowrBGCB9ouLiEg3UYj3kIH9MjhxSD+dLy4iIt1GId6DZpSEWLJ5Lwdr6oIuRUREegGFeA8qLQlR3+hYuL4y6FJERKQXUIj3oFNH9Cc7LVmXJhURkW4R0xA3s4vMbLWZrTOzO6PMzzOzF81suZmtMLMvxLKeoKWlJHHW2ELeWFOOczrVTEREjk/MQtzMkoH7gYuBE4HZZnZixGJfAVY65yYDM4GfmllarGqKBzNKQpTtPcyGiqqgSxERkQQXy5b4dGCdc26Dc64WmAdcHrGMA3LNzIAcYA9QH8OaAlc6LgToqmYiInL8YhniRcDWsMdl/rRwvwQmANuBD4F/cs41Rq7IzG4zs8Vmtri8PLHDb3hBFqMLs3W+uIiIHLdYhrhFmRa5I/hTwDJgKDAF+KWZ9TvqSc496Jyb5pybFgqFurvOHjejJMS7GyqpqWsIuhQREUlgsQzxMmBY2ONivBZ3uC8AzznPOmAjcEIMa4oLpSUhauoaeW/jnqBLERGRBBbLEF8EjDOzUf7BarOA+RHLbAHOBzCzQcB4YEMMa4oLp48eQFpKkq5qJiIixyUlVit2ztWb2e3Aa0Ay8Bvn3Aoz+5I//wHgP4A5ZvYhXvf7vzrnKmJVU7zISkvh9FEDdL64iIgcl5iFOIBz7hXglYhpD4Td3w5cGMsa4lVpSYh7Xl7Ftn2HKcrPDLocERFJQBqxLSAzSrwD9NQaFxGRY6UQD8i4gTkMycvQ+eIiInLMFOIBMTNKS0K8va6CuoajTo0XERHpkEI8QKUlIQ4eqWfZ1n1BlyIiIglIIR6gs8YWkpxk6lIXEZFjohAPUF5mKlOH5et8cREROSYK8YCVloT4cNt+Kg4dCboUERFJMArxgJWO9041e2ttrx/jRkREuplCPGATh+YxIDtNXeoiItJlCvGAJSUZ544rZMGachobIy/yJiIi0jaFeBwoLQlRWVXLiu0Hgi5FREQSiEI8Dpw7zh+Cda261EVEpPMU4nEglJvOxKJ+Ol9cRES6RCEeJ2aMC7Fky14O1NQFXYqIiCQIhXicKC0J0dDoWLhOp5qJiEjnKMTjxCkj+pOTnsIbaxTiIiLSOQrxOJGanMTZYwtYsKYc53SqmYiIdEwhHkdmlITYtu8w68sPBV2KiIgkAIV4HJnhn2r2Nx2lLiIinaAQjyPDBmQxJpTNAo2jLiIinaAQjzOlJQP5+4ZKauoagi5FRETinEI8zswoKeRIfSPvbqgMuhQREYlzCvE4c8boAtJTknRVMxER6ZBCPM5kpCZz+mjvVDMREZH2KMTjUGlJiPXlVWzdUx10KSIiEscU4nGotKQQ0FXNRESkfQrxODQmlENRfqauaiYiIu1SiMchM2NGSYiF6yupa2gMuhwREYlT7Ya4mV1mZiPCHn/PzJab2XwzGxX78vqu0pIQh47U8/7mvUGXIiIicaqjlvgPgHIAM7sU+DzwRWA+8EBsS+vbzhpbQHKS6VQzERFpU0ch7pxzTYdIXwU87Jxb4px7CAjFtrS+rV9GKqcO768QFxGRNnUU4mZmOWaWBJwPvB42LyN2ZQlA6fgQK7YfoPzgkaBLERGRONRRiP8cWAYsBlY55xYDmNlUYEdMKxNKS7zOjjd1qpmIiETRbog7534DlAK3AJeEzdoBfCGGdQlw4pB+FOakqUtdRESiSmlvpn9k+j7n3Db/8SeAK4DNwC9jXl0fl5RknDsuxN9W76ah0ZGcZEGXJCIicaSj7vSngGwAM5sCPA1sASYDv4ppZQJ4Xep7q+v4aNv+oEsREZE4025LHMh0zm33738e+I1z7qf+gW7LYlqZAHDuuELMYMGaciYPyw+6HBERiSMdHp0edv88/KPTnXMaRqyHFOSkc3JRnvaLi4jIUToK8b+Y2VNm9gugP/AXADMbAtTGujjxzBgX4v0te9lfXRd0KSIiEkc6CvF/Bp4DNgHnOOeaUmQw8J3YlSXhSseHaHTw9vqKoEsREZE40tEpZs45Nw94AZhqZp82s9HOuaXOudd6pEJh6rB8cjNSWKAudRERCdPRKWb9gIeAU4HlePvIJ5vZEuAW59yB2JcoKclJnDO2kDfWlOOcw0ynmomISMfd6fcCK4FxzrmrnHNXAmOAD9F54j1qRkmIHftrWLv7UNCliIhInOgoxM92zt0dfjS638X+78CZsS1Nws3wh2B9Y7W61EVExNOVU8wkQEX5mYwbmMMCjaMuIiK+jkL8bTP7nkXshDWzu4B3Y1eWRFNaEuLvG/ZQXVsfdCkiIhIHOgrxrwInA+vM7Fkze8bM1uMNu3p7zKuTVmaUhKhtaOTvG/YEXYqIiMSBjk4xO+Ccuxa4EJgD/A640Dl3DbqKWY+bPmoAGalJGr1NRESAjlviADjn1jvnXnTOzXfOrfcnf72j55nZRWa22szWmdmdbSwz08yWmdkKM3ujC7X3ORmpyZwxukDni4uICNDJEG9Duwe9mVkycD9wMXAiMNvMToxYJh/vamifcc6dBFx7HPX0CaUlITZUVLGpoiroUkREJGDHE+Kug/nTgXXOuQ3OuVpgHnB5xDKfBZ5zzm0BcM7tPo56+oRPThhEarJx02/fY+V2jbUjItKXtRviZnbQzA5EuR0Ehnaw7iJga9jjMn9auBKgv5n9zcyWmNmNXX4FfcywAVnM/YczqKlr4Mpfvc3Ti7d2/CQREemVOjqwLdc51y/KLdc519G1yKN1t0e23lPwhnT9NPAp4C4zKzlqRWa3mdliM1tcXq79wdNGDuDlO87l1BH9+cYzH3Dnsx9QU9cQdFkiItLDjqc7vSNlwLCwx8XA9ijL/ME5V+WcqwAW4J2+1opz7kHn3DTn3LRQKBSzghNJYU46j95yOl/5xBjmLdrK1f+7kC2V1UGXJSIiPSiWIb4IGGdmo8wsDZgFzI9Y5vfAuWaWYmZZwOnAqhjW1KskJxnf+NQJPHzTNLbuqebS+97kzyt3BV2WiIj0kJiFuHOuHm9AmNfwgvkp59wKM/uSmX3JX2YV8AfgA+A94CHn3Eexqqm3On/CIF6+41yGF2Rx6+8W8+M/fEx9Q2PHTxQRkYRmznV0kHl8mTZtmlu8eHHQZcSlmroGvv/iCua+t5UzRxdw7+yphHLTgy5LRESOk5ktcc5Ni5wey+506WEZqcn811WT+O9rJvH+lr1cet+bLN6kIVpFRHorhXgvdO20YTz/5bPJSE1m1oPv8tCbG0i0HhcREemYQryXOnFoP1786jmcd8JA7nl5FV954n0O1tQFXZaIiHQjhXgv1i8jlV/fcCrfvuQEXluxi8t/+Tardx4MuiwREekmCvFezsy4bcYYnrj1dA4eqeeK+9/m+aVlQZclIiLdQCHeR5w+uoCXv3oOJxfn8bUnl/PdFz7kSL1GeRMRSWQK8T5kYL8Mnrj1dP7PjNE89u4WrnvgHcr2apQ3EZFEpRDvY1KSk/jWJRP49Q2nsqG8ikvve4u/rtbF40REEpFCvI/61EmDefGr5zC4XwZfnLOIn/1xNQ2NOg1NRCSRKMT7sJGF2Tz/5bO5+pRi7v3LOm7+7XvsqaoNuiwREekkhXgfl5mWzH9fM4kfXnUyf9+4h0/f+ybvb9kbdFkiItIJCnHBzJg1fTjP/eNZpCQb1//6Hea8vVGjvImIxDmFuDSbWJTHS7efS2lJiLtfXMkd85ZRdaQ+6LJERKQNCnFpJS8rlQdvmMY3PjWelz/YzuX3v8263RrlTUQkHinE5ShJScZXPjGWx245nb1VtXzml28zf/n2oMsSEZEICnFp01ljC3n5jnOZMKQfd8xdyt3zV1Bb3xh0WSIi4lOIS7sG52Uw77YzuOWcUcxZuInrH3yH7fsOB12WiIigEJdOSE1O4q5LT+RXnzuFtbsOcel9b/Hm2vKgyxIR6fMU4tJpl5w8hN/ffjaFOWnc+Jv3uPf1tTRqlDcRkcAoxKVLxoRyeOErZ3PFlCJ+9qc1fPGRRezVKG8iIoFQiEuXZaWl8LPrJvMfV0xk4bpKLr3vLZZv3Rd0WSIifY5CXI6JmXHDGSN4+ktnAnDtA+/w2LubNcqbiEgPUojLcZk8LJ+XvnoOZ44p4LsvfMTXn1pOda1GeRMR6QkKcTlu/bPT+O3Np/H1C0p4Ydk2rrx/IevLDwVdlohIr6cQl26RlGTccf44HvnCdHYfrOHyX77NKx/uCLosEZFeTSEu3WpGSYiX7jiXsQNz+PLj7/MfL62krkGjvImIxIJCXLpdUX4mT/2fM7npzBE8/NZGZj/4Ljv31wRdlohIr6MQl5hIS0ni+5dP5N7ZU1m54wCX3vcmj76zif2H64IuTUSk17BEOyVo2rRpbvHixUGXIV2wdtdB/uXp5XxQtp+M1CQuOXkIs04bzmkj+2NmQZcnIhL3zGyJc27aUdMV4tITnHN8tO0AcxdtYf6y7Rw6Us/oUDazThvGVacUU5iTHnSJIiJxSyEucaO6tp6XPtjBk4u2smTzXlKTjQtOHMT1pw3n3LGFJCWpdS4iEk4hLnFp7a6DzFu0lefeL2NvdR1F+ZlcN20Y104rZmh+ZtDliYjEBYW4xLUj9Q38aeUu5r23lbfWVZBkUFoS4vrThnP+hIGkJusYTBHpuxTikjC27qnmyUVbeXrJVnYdOEJhTjrXnFrM9acNY1RhdtDliYj0OIW4JJz6hkb+trqceYu28tfVu2lodJw+agCzpw/noomDyUhNDrpEEZEeoRCXhLbrQA3PLCnjyUVb2bKnmn4ZKVw5tYhZ04czYUi/oMsTEYkphbj0Co2Njnc3VDJ30VZe+2gntQ2NTC7OY9b04Vw2eSg56SlBlygi0u0U4tLr7K2q5fml25i3aAtrdh0iKy2ZSycNYdb04Uwdlq+BZESk11CIS6/lnGPp1n08+d5WXvxgO9W1DYwflMv1pw3jyqlF9M9OC7pEEZHjohCXPuFgTR0vfbCDee9tYXnZftKSk/jUxMHMPm0YZ4wu0EAyIpKQFOLS56zcfoCnFnsDyRyoqWdEQZY3kMypxQzslxF0eSIinaYQlz6rpq6BP3y0k3mLtvDuhj0kJxmfGD+Q2dOHUVoSIkUDyYhInGsrxHUor/R6GanJXDG1iCumFrGxooonF23lmSVl/HnVLgb1S+faU4dx/WnDGDYgK+hSRUS6RC1x6ZPqGhp5fdVu5i3awhtrynEOzhlbyPWnDePCkwaRnqKBZEQkfqg7XaQN2/cd5unFZTy1eCvb9h2mf1YqV51SzKzThjFuUG7Q5YmIKMRFOtLQ6HhrXQVPLtrCH1fsor7RMXZgDmeNKeCsMYWcObqAvKzUoMsUkT5IIS7SBRWHjvDC0m28ubaC9zbu4XBdA2YwcWgeZ431Qv20kf3JStNhJSISewpxkWNUW9/Isq37WLi+goXrKlm6dS91DY7UZGPqsP7NoT5lWD5pKTrSXUS6n0JcpJtU19azaNPe5lD/aPt+nIPM1GROGzWAs/3u9xOH9iNZg8uISDcIJMTN7CLgF0Ay8JBz7odtLHca8C5wvXPumfbW2a0hfnAnvPR1uPR/IHdQ96xT+pz91XW8s6GSd9ZX8Pb6StbtPgRAXmYqZ4wewNljCzlrTAFjQjkaz11EjkmPnyduZsnA/cAFQBmwyMzmO+dWRlnuR8BrsaqlTXs3wYa/wpxPw00vQr8hPV6CJL68rFQumjiYiyYOBmD3gRoWrq/k7XUVLFxfyWsrdgEwqF+6d4DcmALOHltIUX5mkGWLSC8Qs5a4mZ0J3O2c+5T/+FsAzrn/iljun4E64DTgpR5tiQNsXgiPXQO5g+Hml6Df0O5bt/R5zjm27KluDvV31ldSWVULwIiCLM4aU+gf/V5AQU56wNWKSLwKYsS2ImBr2OMy4PSIooqAK4Hz8EI8KjO7DbgNYPjw4d1b5Yiz4IbnvCD/7SVeizx/WPduQ/osM2NEQTYjCrKZPX04zjlW7zrIwnWVLFxfwUvLtzP3vS0AnDA4l7PGFHL22AKmjxpAboZOZxOR9sWyJX4t8Cnn3K3+4xuA6c65r4Yt8zTwU+fcu2Y2hyBa4k22LoLHroLMfLjpJeg/ovu3IRKhvqGRD7ftZ+F6L9QXb9rLkfpGkpOMScV5nDWmgLPHFHLKiP5kpGoUOZG+qscPbOtMd7qZbQSajvQpBKqB25xzL7S13pgenb5tCTx6JaT381rkA0bFZjsibaipa+D9LXubW+rLy/bT0OhIS0li2oj+Xtf72EImFeXpwi0ifUgQIZ4CrAHOB7YBi4DPOudWtLH8HIJsiTfZvgx+dzmk5cBN86FgTOy2JdKBgzV1LNq0h7fXVbJwfSWrdhwAICc9hdNHDWg+SG78oFxdK12kF+vxfeLOuXozux3vqPNk4DfOuRVm9iV//gOx2vZxGTrFa4X/7nKYc6l3v3Bs0FVJH5Wbkcp5JwzivBO8UyArDx3h3Q17eHu9d5Dc6x/vBqAgO40z/K73ScV5jA5lazQ5kT5Ag720ZdcKeOQzkJTiBXmoJPbbFOmibfsO8876Shauq+Dt9RXsOnCkeV5RfiZjBuYwJpTN2IE5jAl5t8KcNJ2vLpJgNGLbsdi9Ch65DDAvyAee0DPbFTkGzjk2VlTx8c6DrNt9iPXl/m13FYfrGpqX65eR0hLq/s+xA3MY1j9T+9lF4pRC/FiVr/aCvLHBC/JBJ/bctkW6QWOjY8eBGtaHBbsX8lWUH2xpuacmGyMLwlrtA7ObW+/Z6eqaFwmSQvx4VKz1gryhFm78PQw+uWe3LxIj+w/XsSEs1JtCfnNlNQ2NLf8bhuRl+IGe3aoVPzA3XV3zIj1AIX68Ktd7QV5X7QX5kMk9X4NID6mtb2TLnirW7faDvbkVX8WhI/XNy+WmpzDa3+/e1GofOzCHEQVZpKprXqTbKMS7w56NXpAfOQA3vABFpwRTh0hAnHPsPnikZZ+734Jft/sQOw/UNC+XkmQML8hqDvWmVvyYgTn000h0Il2mEO8uezfDI5fC4f3ecK3FR72nIn3SoSP1bAjf5+634jdVVlHX0PJ/ZmBuevM+97FhB9cNyctQ17xIGxTi3WnfFq9FXlUJn38Whp/e8XNE+qj6hka27Klu2ee++xDr/KA/WNPSNZ+Vlhx1v/vIgmzSUtQ1L32bQry77S/zgvzQbvjcMzDizKArEkkozjkqDtW2Oh1u3e5DbCivYtu+w83LJScZwwdktTpivink8zLVNS99g0I8Fg5s94L8wA743FMw8pygKxLpFapr69ngt9xb9r9XsbGiitqGxublQrnpzQfVhbfeh6prXnoZhXisHNzpjey2bwt89kkYXRp0RSK9Vn1DI2V7Dx/Vel+3+xAHIrrmR4f8fe5+sDcdNZ+eoqvBSeJRiMfSod1ekO/dCLPnwpjzgq5IpE9xzlFZ1dI133ze++5DbXTNZ7cesS6UQ16WuuYlfinEY62qwgvyynUw6wkY98mgKxIRWnfNh58SF9k1X5iTftRBdWMH5jCkX4auECeBU4j3hKpKePRyb6jW6x+HkguDrkhE2tDQ6CjbW91qn3vTUfP7D9c1L5eZmtxqCNqmkB9ZqK556TkK8Z5SvQcevQJ2rYTrH4XxFwddkYh0QVPXfHirvamLPrxr3gyG5mUysjCLkQXZjCrMZmRBNiMLsxg2QAEv3Ush3pMO74VHr4KdH8K1v4UJlwVdkYh0g8O1DWyoONTcHb+poopNldVsqqxiX3VL6z3JYGh+ZnOojyxoCvhshg/I0nnv0mUK8Z5Wsx8euxq2L4WrH4aTrgi6IhGJoX3VtV6gV3inwm2urGKj/zi8e74p4Jta7iMKsrz7hdkM66+Al+jaCnFdXzBWMvLg88/B49fAM18E1wATrw66KhGJkfysNKZkpTFlWP5R8/ZV13ot98oqNlVU+z+r+P2yba1OjUsyKOqf2arlPqowixEFCniJTiEeSxn9vGFZH78Onr3Vuyb5pOuCrkpEelh+VhpTh6cxdXj/VtOdc+yrrmNjZVjXvB/2Lyzb1mpY2iSD4v5ZLS33sK76YQN01bi+SiEea+m58Pln4Inr4bnboLEepnw26KpEJA6YGf2z0+ifncYpUQJ+b3Vdc9f8pgqve35zZRXPL20d8MlJRlF+ptdyL/Ba7k1d9MX9MxXwvZhCvCekZcNnn4J5s+GFL3st8lNuCLoqEYljZsaA7DQGZKdx6oijA35PVW2rlnvT/aWb93LwSOuAL/a76IfkZZCRmuzfkryfKUlkpnnT0lOSvfsp3jzvvrdsemoymanJpCabhrSNIwrxnpKWBbPnwbzPwvzbvRb5tC8EXZWIJCAzoyAnnYKc9HYCvoqNFV7LvWl//KodB6ipa6CmvpHa+sY21t6+JMMLeP/LQHpqUvP9jNQkL/TDwj8z8ktD2OO25mWGrUsD7bRPId6TUjNh1lx48vPw0j97B7uddmvQVYlIL9I64Ae0uVxDo+NIfQM1dY3U1DVwuK7BC3j/cdP9lumt5x9uul/fQE1tg/ezrpG9VXXU1DdwJGK9jcd4IlRGahL9s9K8W3Yq/bO83on8rDQGZKV6uyMi5melJfeZ3gKFeE9LzYBZj8NTN8LL/+J1rZ/+f4KuSkT6mOQkIysthay02G/LOUddg+NwXQNHonw5aPpCcKS+gcO1Dc29BYdrG6g6Us/e6jr2Vdeyp7qW7fsOsKeqttVpe5HSUpLon5XaHO5e6Ke2hH92qv8loCX8c9JTEjL4FeJBSEmH6x6Fp2+GV7/pda2f+ZWgqxIRiQkzIy3FvFPkuuka8PUNjew/XMfe6jr2Vteyt6rW+1ld13x/T5UX/qt2HmCf/0WgrR6B1GQjPyutOfzDAz+ype/dT6NfRvDBrxAPSkoaXPeIdw75a9/2WuRn3xF0VSIiCSElOal5t0FnNTY6DtTUsaeqddhHhv/eqjrW7T7UPL2hjeRPTjL6Z7W06sNb+189byzZ6bGPWIV4kJJT4ZrfwHP/AH+6y2uRn/v1oKsSEemVkpK81nZ+F/YhOOc4UFPvdedX1bKvuulLwNHhv7mymqVb97GvupZ/On9cDF9JC4V40JJT4aqHwJLh9e97LfLSbwRdlYiI4O0KyMtMJS8zlREF2Z16jnOux7rZFeLxIDkFrnoQklLgr/d4LfKZd3qXSRIRkYTSk/vJFeLxIikZrviV9/ONH3pBft53FeQiItImhXg8SUqGz/zS+/nmT7wg/+TdCnIREYlKIR5vkpLg0l94Xetv/9wL8gvvUZCLiMhRFOLxKCkJPv0zL8jf+aV3sNtF/6UgFxGRVhTi8coMLv6xd9T63//Xa5Ff8t8KchERaaYQj2dmXgs8KdlrkbsGuOSnXktdRET6PIV4vDPz9omH7yO/9BcKchERUYgnBDPvKPWkFP+o9Qbv9LPcIepeFxHpwxTiicLMC+6kFO888mWPQ3oehMbDwBMgNMG/P0HhLiLSRyjEE4kZfOJbMPaTsGMZ7F4F5R/Dqpfg/d+1LNcq3P2bwl1EpNdRiCeiYad5t3CHyr1AL//YD/fV8PHL0cO9qcXeFPD9hircRUQSkEK8t8gJebdR57aeXlXR0mIv/xh2fwyrX4Glj7Ysk97PD3e/xR4a73XPK9xFROKaQry3yy70gr1T4f5q2+EeOqFl37vCXUQkLijE+6r2wr25S/5jr1u+zXD3W+xN+977FSncRUR6kEJcWssuhOxzYOQ5rae3CvfV3v3Vf4Clj7Usk5YbcbS833pXuIuIxIRCXDqno3Bv6pIv/xjWvBY93AefDEOnwJDJMPBESEnv0ZcgItLbKMTl+LQZ7pV+uK/ywn33KvjoOVjyW29+UioMOtEL9CFTvNugkyA1o6dfgYhIwlKIS2xkF0D22TDy7JZpzsHejbB9GexY7p3rvnJ+y2lwSSleN/yQyS0t9kETIS0rgBcgIhL/FOLSc8xgwGjvNvEqb5pzsG+LF+g7lnsBv+ZVWOZ3x1sSFI73Q32KF+yDT4b0nGBeg4hIHFGIS7DMoP8I73bi5d405+DAttYt9nWvw/K5TU+CwnFeqDe12AdPgox+gbwEEZGgxDTEzewi4BdAMvCQc+6HEfM/B/yr//AQ8I/OueWxrEkSgBnkFXu3CZe2TD+wo3WLfdNb8OFTLfMHjGndYh8yGTLze7R0EZGeFLMQN7Nk4H7gAqAMWGRm851zK8MW2wiUOuf2mtnFwIPA6bGqSRJcvyHebfzFLdMO7W4J9R3LYOt78NGzLfP7j2zdYh8yBbIG9GTVIiIxE8uW+HRgnXNuA4CZzQMuB5pD3Dm3MGz5d4HiGNYjvVHOQBh3gXdrUlXR0g2/fRlsXworX2iZnzcchvot9SFTvYDPLuzZukVEukEsQ7wI2Br2uIz2W9m3AK/GsB7pK7ILYez53q1J9R7Y+UHr/eyrXmyZ36+opRu+qUs+d1CPli0i0lWxDPFoQ3S5qAuafQIvxM9pY/5twG0Aw4cP7676pC/JGgCjZ3q3Jof3wc4PW+9nX/0KzX+mOYOhYIwX8P2Gevvow+9nFWgkOhEJVCxDvAwYFva4GNgeuZCZTQIeAi52zlVGW5Fz7kG8/eVMmzYt6hcBkS7LzD96/PgjB71gb2qx79sCW9/1DqprrGv9/OT0KOFe5N8v8qZn9lfQi0jMxDLEFwHjzGwUsA2YBXw2fAEzGw48B9zgnFsTw1pEOic9F0ac5d3CNTZCVTkcKIP92+DA9tb3N7/t/XQNrZ+Xkhkl3CPuZ+T33aB3DuqqvS9EyTrjVaSrYvapcc7Vm9ntwGt4p5j9xjm3wsy+5M9/APgeUAD8yrx/YvXOuWmxqknkmCUlefvIcwdB0anRl2ls8I6WP7DNu+3f1vr+xgVwcAe4xtbPS81qu8u+6X5GXnwEvXNQXwNHDkHtQf/noZaf4fePHOzgsf8T5w3DO2AUFIxtuRWO835mh+LjtYvEIXMusXqnp02b5hYvXhx0GSLHpqEeDu1qO+gPbIODOznq8JG0nIgu++KjW/htDXZTXxslRJsCuKoLges/J7K3oS0pGV7d6TneRXDSc8Ie53i9Hmk5kJYNNfuhci1UrIM9G6DhSMt60vO8YxOaQr35NsZ7rkgfYGZLojVy1X8l0pOSU7zgzStqe5mGOi/I2wr63Su9Fn9k0Kf384Ldklu3khtqO1dbUmpLwDaFbXqud27+USHc0eMcSE49tveosQH2b4XKdV6oV67zAn7zQvjgydbL9is6uuVeMBbyh0NS8rFtXySBKMRF4k1yKuQP825tqa/1uuYPbPfDvazlvnNdDFz/cbxcGjYp2Rukp/9IGPvJ1vNqq72WeuXa1iH/0TNea75Jcpo3Rn+07nmdVSC9iEJcJBGlpLWMOd+XpGXB4IneLZxzUF0JFWtbWu6V673Ha15rfWZBRn5YqI+BAj/cB4zWFfMk4SjERSTxmfnXti+EEWe2ntdQD/u3tIR6U8hvXBB2UR1f3rCWYA8P+bxidc9LXFKIi0jvlpzScgnc8OF5wTtuoLl7PizkP3gSjhwIW0e69/zCpu75cRA6AQZOUOtdAqUQF5G+Kz0HhkzybuGc8w4ebO6a9/e/7/4YVr8KjfXecpbkBfqQSd517gefDIMnQ3ZBz78W6ZMU4iIikcxaxgUYeXbreQ11sHczlK/yh+39ADa/Ax8+3bJM7lAv0MPDPX+kN96ASDdSiIuIdEVyqtetXjgWJlzWMr3pIjs7P2wJ93V/bjmvPr0fDJrYEupDJnld8vFyVoAkJIW4iEh3iHaRnbrDsNtvsTcF/NLHoK7Km5+U4gX54PDu+InemPsinaAQFxGJldRMKDrFuzVpbPQOpgtvta9/HZY/0bJM/vCwYPd/5hXr/HY5ikJcRKQnJSW1dMdPvKpl+sFdsMvvhm8K949fpnlkvsz+rUN98MlQWHLsI+NJr6AQFxGJB00H0oWPUnfkkDfMblOrfccHsOgh7yI04J36NnCCv499svdz0EnecLkSG855QwO7Bu8shcZ673Fj+ON6yB/RIwcyKsRFROJVeg4Mm+7dmjTUe6e87fygJdw/fhmWPtqyzIDRrbvjh0yCnEHH3h3vnHf1vfCQigytqGHW1jJhj49ab7TnRSzjom0ncr0dLNMcwm38bGt+Zy8AdOfWti9K1I0U4iIiiSQ5BQae4N0mXedNc84bO7+pG37nctixDFa+0PK87JA3Il1zuLUXtlECMR5YsjdyXlKKf0v2p4U9bnd+ijdkcVJWy2NLaj0/ch3Rthn+05KjT++hsw4U4iIiic6s5ep44y9qmV6zH3Z+1HJ0/MGd3j70NgMwMszaCbeoy7QViikdrKutZSJCUgf2HUUhLiLSW2XkeYPVRA5YI72Ghg8SERFJUApxERGRBKUQFxERSVAKcRERkQSlEBcREUlQCnEREZEEpRAXERFJUApxERGRBKUQFxERSVAKcRERkQSlEBcREUlQCnEREZEEpRAXERFJUOacC7qGLjGzcmBz0HXEgUKgIugi+gC9zz1D73PP0PvcM2LxPo9wzoUiJyZciIvHzBY756YFXUdvp/e5Z+h97hl6n3tGT77P6k4XERFJUApxERGRBKUQT1wPBl1AH6H3uWfofe4Zep97Ro+9z9onLiIikqDUEhcREUlQCvEEYmbDzOyvZrbKzFaY2T8FXVNvZmbJZrbUzF4KupbezMzyzewZM/vY/9s+M+iaeiMz+5r/f+MjM5trZhlB19QbmNlvzGy3mX0UNm2Amf3JzNb6P/vHavsK8cRSD/yLc24CcAbwFTM7MeCaerN/AlYFXUQf8AvgD865E4DJ6D3vdmZWBNwBTHPOTQSSgVnBVtVrzAEuiph2J/C6c24c8Lr/OCYU4gnEObfDOfe+f/8g3j+7omCr6p3MrBj4NPBQ0LX0ZmbWD5gBPAzgnKt1zu0LtKjeKwXINLMUIAvYHnA9vYJzbgGwJ2Ly5cAj/v1HgCtitX2FeIIys5HAVODvAZfSW/0c+CbQGHAdvd1ooBz4rb/r4iEzyw66qN7GObcN+AmwBdgB7HfO/THYqnq1Qc65HeA1voCBsdqQQjwBmVkO8Czwz865A0HX09uY2aXAbufckqBr6QNSgFOA/3XOTQWqiGHXY1/l75O9HBgFDAWyzezzwVYl3UEhnmDMLBUvwB93zj0XdD291NnAZ8xsEzAPOM/MHgu2pF6rDChzzjX1KD2DF+rSvT4JbHTOlTvn6oDngLMCrqk322VmQwD8n7tjtSGFeAIxM8Pbd7jKOfezoOvprZxz33LOFTvnRuId/PMX55xaLTHgnNsJbDWz8f6k84GVAZbUW20BzjCzLP//yPnoAMJYmg/c5N+/Cfh9rDaUEqsVS0ycDdwAfGhmy/xp33bOvRJcSSLH7avA42aWBmwAvhBwPb2Oc+7vZvYM8D7eWS5L0eht3cLM5gIzgUIzKwP+Dfgh8JSZ3YL3BeramG1fI7aJiIgkJnWni4iIJCiFuIiISIJSiIuIiCQohbiIiEiCUoiLiIgkKIW4SB9hZg1mtizsdqc//W9mttrMlpvZ203nbJtZmpn93MzW+1dj+r0/pnzT+gab2Tx//koze8XMSsxsZPgVnfxl7zaz/9uzr1ik99N54iJ9x2Hn3JQ25n3OObfYzG4D/hv4DPCfQC5Q4pxrMLMvAM+Z2en+c54HHnHOzQIwsynAIGBrDF+DiIRRiItIuAXAP5tZFt6gK6Occw0AzrnfmtkXgfMAB9Q55x5oeqJzbhk0X5xHRHqAQlyk78gMG+kP4L+cc09GLHMZ8CEwFtgS5QI7i4GT/PvtXSBmTMS2BuNdRUtEupFCXKTvaK87/XEzOwxswhsGdQBeazuS+dOtg22tD9+Wmd3dxVpFpBMU4iIC/j7xpgdmtgcYYWa5zrmDYcudArzo37+mJwsUkaPp6HQROYpzrgp4BPiZmSUDmNmNQBbwF/+Wbmb/0PQcMzvNzEqDqFekr1KIi/QdmRGnmP2wg+W/BdQAa8xsLd6VmK50PuBK4AL/FLMVwN3A9li+ABFpTVcxExERSVBqiYuIiCQohbiIiEiCUoiLiIgkKIW4iIhIglKIi4iIJCiFuIiISIJSiIuIiCQohbiIiEiC+v/xgrvtEDEnIgAAAABJRU5ErkJggg==",
      "text/plain": [
       "<Figure size 576x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#Train Dropout Model\n",
    "train_model(learner = Maxout_D_learner, n_epochs = 10, train_iters = 750, valid_iters = 50)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Path('models/Maxout_D_Model.pth')"
      ]
     },
     "execution_count": 31,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Save Maxout model (with bias)\n",
    "Maxout_D_learner.save(\"Maxout_D_Model\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### COMPARING PERFORMANCE + MODEL VALIDATION."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Load all models (learner object is being created again to avoid running all previous training and lr_find cells)\n",
    "#Declare Learners dynamically\n",
    "def declareLearners(model):\n",
    "    globals()[model + \"_learner\"] = Learner(training_dataloader, globals()[model + \"_Model\"], metrics = accuracy)\n",
    "    globals()[model + \"_learner\"].load(model + \"_Model\")\n",
    "\n",
    "[declareLearners(model) for model in model_list]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Due to IPython and Windows limitation, python multiprocessing isn't available now.\n",
      "So `number_workers` is changed to 0 to avoid getting stuck\n"
     ]
    }
   ],
   "source": [
    "#Define new datablock for validation data - to create a new dataloader for testing\n",
    "#Set splitter and transforms to None as we do not want to split or augment the dataset\n",
    "\n",
    "validationDatablock = DataBlock(\n",
    "            blocks = (ImageBlock, CategoryBlock),\n",
    "            get_items = get_image_files,\n",
    "            splitter = None,\n",
    "            get_y = parent_label,\n",
    "            batch_tfms = None)\n",
    "\n",
    "#Create new dataloader -> set shuffle to false and do not drop the last batch\n",
    "valid_data = datablock.dataloaders(dataset/\"testing\", batch_size = 64, shuffle_train = False, drop_last = False)\n",
    "#Use .train to access testing images -> this is simply as the testing images have been shuffled into the .train directory via fastai by default\n",
    "#Use .tee to create 4 ITERATORS - one for testing each model\n",
    "import itertools\n",
    "ReLU_valid_loader, Maxout_WB_valid_loader, Maxout_B_valid_loader, Maxout_D_valid_loader = itertools.tee(valid_data.train, 4)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Evaluate perfromance based on 8000 samples from validation dataset (obtained from training dataloader)\n",
    "#A seperate test set in this case is not required; gradients were off during validation and thus the model did not \"learn\" during those portions\n",
    "\n",
    "#Use sci-kit learn to generate a confusion matrix and seaborn to plot \n",
    "from sklearn.metrics import confusion_matrix\n",
    "import seaborn as sns\n",
    "\n",
    "#Declare true and prediction arrays for all three models\n",
    "\n",
    "TRUE_AND_PREDS = {}\n",
    "def assignValidationArrays(model):\n",
    "    TRUE_AND_PREDS[model] = {\"y_true\" : [], \"y_preds\" : []}\n",
    "\n",
    "#Populate True and Preds dictionary\n",
    "[assignValidationArrays(m) for m in model_list]\n",
    "\n",
    "#TensorImages (found in the dataloader) are a custom FastAi datatype that have no equivalent to .index - this functionality is needed to extract categories from predictions\n",
    "#This function simply iterates and returns the needed index; as it will only be iterating over 10 samples this is sufficient \n",
    "def findTensorImageIndex1D(tensor, val):\n",
    "    for i, obj in enumerate(tensor):\n",
    "        if obj == val:\n",
    "            return i\n",
    "\n",
    "#Generate prediction and label array - for determining accuracy and plotting confusion matrix\n",
    "def generatePredictions(learner, y_true, y_preds, validation_data):\n",
    "    #This loop iterates over each batch, feeds to the model, and generates labels and predictions\n",
    "    #As we require 8000 validation samples per learner, this loop must be executed (8000/64) 125 times\n",
    "    for i in range(125):\n",
    "        #Gets one BATCH (size 64) of inputs and labels\n",
    "        inputs, labels = next(validation_data)\n",
    "        #Add labels to y_true\n",
    "        y_true.extend([*labels])\n",
    "        #Passes one batch into Maxout model and obtains predictions\n",
    "        predictions = learner.model(inputs)\n",
    "        #Converts predictions into categories -> predictions are continous and thus must be processed to get categories\n",
    "        for pred_distribution in predictions:\n",
    "            pred_class = findTensorImageIndex1D(pred_distribution, max(pred_distribution))\n",
    "            y_preds.append(pred_class)\n",
    "\n",
    "#Generate predictions for each model\n",
    "for model in model_list:\n",
    "    generatePredictions(learner = globals()[model + \"_learner\"], y_true = TRUE_AND_PREDS[model][\"y_true\"], y_preds = TRUE_AND_PREDS[model][\"y_preds\"],\n",
    "                        validation_data = globals()[model + \"_valid_loader\"])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1728x1728 with 8 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#Generate heatmaps and confusion matricies for each model\n",
    "fig, (ReLU_ax, Maxout_WB_ax, Maxout_B_ax, Maxout_D_ax) = plt.subplots(4, 1, figsize = (24, 24))\n",
    "#Heatmap plotter \n",
    "def plotHeatmaps(model):\n",
    "    #Use ScikitLearn to generate a 2D confusion matrix\n",
    "    globals()[model + \"_confusion_matrix\"] = confusion_matrix(TRUE_AND_PREDS[model][\"y_true\"], TRUE_AND_PREDS[model][\"y_preds\"])\n",
    "    #Feed into seaborn to plot heatmap\n",
    "    globals()[model + \"_heatmap\"] = sns.heatmap(globals()[model + \"_confusion_matrix\"], annot = True, fmt = \"g\", ax = globals()[model + \"_ax\"])\n",
    "    #Set chart title, x axes, and y axes\n",
    "    globals()[model + \"_ax\"].set_title(model + \" MODEL CONFUSION MATRIX.\")\n",
    "    globals()[model + \"_ax\"].set_xlabel(\"PREDICTED VALUES\")\n",
    "    globals()[model + \"_ax\"].set_ylabel(\"ACTUAL VALUES\")\n",
    "\n",
    "[plotHeatmaps(m) for m in model_list]\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "RELU MODEL ACCURACY: 0.957\n",
      "CUSTOM MAXOUT MODEL ACCURACY (WITHOUT BIAS): 0.92075\n",
      "CUSTOM MAXOUT MODEL ACCURACY (WITH BIAS): 0.926375\n",
      "CUSTOM MAXOUT MODEL ACCURACY (DROPOUT, WITH BIAS): 0.9095\n"
     ]
    }
   ],
   "source": [
    "#Define accuracy function\n",
    "def calculateAccuracy(labels, preds):\n",
    "    total_correct = 0\n",
    "    #Find number of correct predictions\n",
    "    [total_correct := total_correct + 1 for label, pred in zip(labels, preds) if label == pred]\n",
    "    #Find accuracy\n",
    "    accuracy = total_correct/len(labels)\n",
    "    return accuracy\n",
    "#Define accuracy getter function\n",
    "def getAccuracy(model):\n",
    "    globals()[model + \"_accuracy\"] = calculateAccuracy(TRUE_AND_PREDS[model][\"y_true\"], TRUE_AND_PREDS[model][\"y_preds\"])\n",
    "#Get accuracies of all 4 models\n",
    "[getAccuracy(m) for m in model_list]\n",
    "\n",
    "#Print\n",
    "print(\"RELU MODEL ACCURACY:\", ReLU_accuracy)\n",
    "print(\"CUSTOM MAXOUT MODEL ACCURACY (WITHOUT BIAS):\", Maxout_B_accuracy)\n",
    "print(\"CUSTOM MAXOUT MODEL ACCURACY (WITH BIAS):\", Maxout_WB_accuracy)\n",
    "print(\"CUSTOM MAXOUT MODEL ACCURACY (DROPOUT, WITH BIAS):\", Maxout_D_accuracy)"
   ]
  }
 ],
 "metadata": {
  "interpreter": {
   "hash": "287ad233368e001a2a800d2a7955348d4c82e83267c9fb56d5631ebf5a35ee7c"
  },
  "kernelspec": {
   "display_name": "Python 3.7.13 ('Maxout')",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.10.4"
  },
  "orig_nbformat": 4
 },
 "nbformat": 4,
 "nbformat_minor": 2
}